The Agent Skills Directory

[SAFE]: No security issues detected. The skill follows best practices for a FinOps orchestration task, utilizing local data files and sub-agents to perform specialized analysis.
[DATA_EXFILTRATION]: Accesses local project files containing operational data related to the skill's FinOps purpose.
Evidence: The skill reads out/focus-normalized.csv and resources/sample-billing/utilization-sample.csv to analyze resource usage.
Observation: No sensitive system paths, credentials, or network operations were identified.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external CSV files, creating a potential injection surface.
Ingestion points: out/focus-normalized.csv in Phases 1, 2, and 3.
Boundary markers: None explicitly mentioned in the workflow instructions.
Capability inventory: The skill is restricted to generating markdown reports (.md) via sub-agents and local tools. It lacks high-risk capabilities such as arbitrary command execution or outbound network access.
Sanitization: Not explicitly implemented, but risk is naturally limited by the skill's read-only and report-generating scope.

optimize